Commit Graph

1613 Commits

Author SHA1 Message Date
Jan Tojnar
c5881ec2c9
Merge remote-tracking branch 'upstream/master' into staging 2018-11-30 20:09:45 +01:00
Matthew Bauer
28e2277305 make-derivation: remove selfConsistent check
version is set in lots of places but might not need to be in a name.

Alternative to #50364.
2018-11-30 17:58:33 +01:00
Frederik Rietdijk
1828a5c5ba Merge master into staging-next 2018-11-30 17:46:21 +01:00
Matthew Bauer
e6834171b9
Merge pull request #51183 from matthewbauer/make-derivation-no-assertions
Fix breaking changes to make-derivation
2018-11-29 19:35:55 -06:00
Vaibhav Sagar
b5504e26b0 pkgs/stdenv/generic/make-derivation.nix: s/targetPlatform/hostPlatform/ 2018-11-29 17:54:59 -05:00
Frederik Rietdijk
9b81c7e455 Merge staging-next into staging 2018-11-29 09:18:35 +01:00
Matthew Bauer
a56fe056ec make-derivation: don’t add host suffix if there is no c compiler
Some trivial builders use the name attr to choose the exec name
produced. For example nixos-install,

  nixos-install = makeProg {
    name = "nixos-install";
    src = ./nixos-install.sh;
    nix = config.nix.package.out;
    path = makeBinPath [ nixos-enter ];
  };

When cross compiling, this puts the prog in,

  /bin/nixos-install-powerpc64le-unknown-linux-gnu
2018-11-28 12:44:10 -06:00
Matthew Bauer
d64f288154 make-derivation: don’t assert on separateDebugInfo
When separateDebugInfo = true & !stdenv.hostPlatform.isLinux, we
always gave an error. There is no reason to do this. Instead, just
don’t add separate debug info when we aren’t on Linux.
2018-11-28 12:29:00 -06:00
Matthew Bauer
13e45aeed3 make-derivation: remove selfConsistent check
version is set in lots of places but might not need to be in a name.

Alternative to #50364.
2018-11-28 12:26:21 -06:00
Léo Gaspard
8ecd555ded
Merge branch 'pr-46056' into staging
* pr-46056:
  binutils: use shared libs
  binutils: fix #44936 the huge size regression
2018-11-28 09:55:05 +09:00
Vladimír Čunát
a5de78b7d7
Merge branch 'master' into staging-next 2018-11-26 10:28:00 +01:00
Matthew Bauer
7eeb02d47b Revert "make-derivation: add disallowedReferences in strictDeps"
This reverts commit 8dbfb61e46.
Also reverts commit fc99c337ed.

Fixes #50915
2018-11-25 15:53:28 -06:00
Frederik Rietdijk
c31cb577ae Merge master into staging-next 2018-11-22 09:57:08 +01:00
Matthew Bauer
fc99c337ed make-derivation: don’t disallow propagated native inputs
propagateNativeBuildInputs will end up going in the output derivation.
This case is allowed to end up in references because of that. Sorry
for the disruption!

Fixes #50865
2018-11-21 09:39:47 -06:00
Daiderd Jordan
c9223a17bc
Revert "patch-shebangs: use --build for auto patch shebangs"
Completely breaks darwin. Every package in the stdenv that has shebangs
in the output will end up with references to bootstrap-tools.

This reverts commit eb7c50a993.
2018-11-14 23:37:31 +01:00
Daiderd Jordan
25fafd2eb5
Revert "darwin/stdenv: bash is a build input"
Completely breaks darwin. Every package in the stdenv that has shebangs
in the output will end up with references to bootstrap-tools.

This reverts commit 2f2e635dd5.
2018-11-14 23:37:28 +01:00
Matthew Bauer
f19bb8321a
Merge pull request #35304 from volth/patch-97
[staging] substitute() print warning if does nothing
2018-11-12 12:52:51 -06:00
Frederik Rietdijk
3b052406ea Merge staging-next into staging 2018-11-12 19:01:36 +01:00
Frederik Rietdijk
daf3297cb4 Merge master into staging-next 2018-11-12 18:59:08 +01:00
Léo Gaspard
2986ce16a8
meta.tests: rename into passthru.tests
Nix currently rejects derivations in `meta` values. This works around
that limitation by using `passthru` instead.

Closes https://github.com/NixOS/nixpkgs/issues/50230
2018-11-11 23:11:46 +09:00
Léo Gaspard
cae5598611
meta.tests: drop meta.needsVMSupport
Its job is already handled by `requiredSystemFeatures`
2018-11-11 23:11:46 +09:00
Léo Gaspard
83b27f60ce
tests: split into a separate all-tests.nix file
This will make the list much easier to re-use, eg. for `nixosTests`

The drawback is that this approaches makes the
```
nix-build release.nix -A tests.opensmtpd.x86_64-linux
```
command about twice as slow (3s to 6s): it now has to evaluate `nixpkgs`
once for each architecture, instead of just having the hardcoded list of
tests that allowed to say “ok just evaluate for x86_64-linux”.

On the other hand, complete evaluation of `release.nix` should be much
faster because we no longer import `nixpkgs` for each test: testing with
the following command went from 30s to 18s, and that's just for a few
tests.
```
time nix-instantiate --eval --strict nixos/release.nix -A tests.nat
```
I initially wanted to test on the whole `release.nix`, but there are too
many broken tests and it takes too long to eval them all, especially
compared to the fact that the current implementation breaks some setup.

Given developers can just `nix-build nixos/tests/my-test.nix`, it sounds
like an overall win.
2018-11-11 23:11:46 +09:00
Frederik Rietdijk
1d3bff25db Merge staging-next into staging 2018-11-11 14:28:08 +01:00
Frederik Rietdijk
53d00c3351 Merge master into staging-next 2018-11-10 11:08:54 +01:00
Matthew Bauer
c8aff96110
Merge pull request #49608 from matthewbauer/cross-patch-shebangs-2
Restore cross-patch-shebangs branch
2018-11-07 13:37:02 -06:00
Matthew Bauer
6d531f3541 make-derivation: enable pie hardening with musl
Fixes #49071

On ld.gold, we produce broken executables when linking with the Musl
libc. This appears to be a known bug when using ld.gold and Musl. This
thread describes the workaround as enabling PIE when using ld.gold and
Musl:

https://www.openwall.com/lists/musl/2015/05/01/5

By default we don’t enable PIE to avoid breaking things. But in the
Musl case we are breaking things by not enabling PIE. So this adds a
special case for defaultHardeningFlags which keeps the pie hardening
for everything. Any packages that break with PIE can add the pie flag
to disableHardeningFlags array (a no-op for now on anything but Musl).
2018-11-07 15:24:51 +01:00
Timo Kaufmann
6141939d6e
Merge pull request #44439 from Ekleog/meta-tests
[RFC] Use `meta.tests` to link from packages to the tests that test them
2018-11-07 00:05:22 +01:00
Patrick Hilhorst
320c9c10de
make-derivation: use pname-version as default name if both are present 2018-11-06 00:04:21 +01:00
Frederik Rietdijk
322f87137c Merge master into staging-next 2018-11-04 11:33:17 +01:00
Frederik Rietdijk
abea6f461a Revert "Merge pull request #49398 from Synthetica9/implement-rfc0035" to fix eval
This reverts commit 3fc7d5eb83, reversing
changes made to 1fddf2b689.

The idea is good, however, before enforcing, make sure all occurences
are fixed.
2018-11-04 11:02:43 +01:00
Frederik Rietdijk
cb4ff927a1 Merge master into staging-next 2018-11-04 08:49:24 +01:00
Jörg Thalheim
3fc7d5eb83
Merge pull request #49398 from Synthetica9/implement-rfc0035
Implement rfc0035: default `name` from `pname`
2018-11-03 21:33:36 +00:00
Matthew Bauer
8dbfb61e46 make-derivation: add disallowedReferences in strictDeps
When strictDeps = true, we don’t want native build inputs to end up in
the output. For instance gcc is a builtin native build input and
should only show up in an output if it is also listed in buildInputs.

/cc @ericson2314
2018-11-02 19:31:51 -05:00
Matthew Bauer
2f2e635dd5 darwin/stdenv: bash is a build input
patch shebangs needs to be in build inputs for it to get into
HOST_PATH.
2018-11-02 00:27:14 -05:00
Matthew Bauer
eb7c50a993 patch-shebangs: use --build for auto patch shebangs
In strictDeps=false, autoPatchshebangs should use
--build (corresponding to PATH) to lookup commands. This restores the
previous behavior of patchshebangs so that we don’t break stuff that
isn’t careful in the buildInputs vs. nativeBuildInputs distinction.
Unfortunately this won’t work under cross compilation.
2018-11-02 00:27:14 -05:00
John Ericson
e3082c313b Merge remote-tracking branch 'upstream/master' into release-lib-cleanup 2018-11-01 16:47:42 -04:00
John Ericson
f2ed7c7af9 linux bootstrap tools: Use right system for some raw derivations
This allows cross builds to work. Evidentallyy this has been done wrong
since I combined the bootstrap tool creation files in
ab651d2c9b. Oops!
2018-11-01 16:22:00 -04:00
John Ericson
ffaffb36d1 linux bootstrap-tools: use stdenv.*Platform to avoid deprecation warning 2018-11-01 16:18:51 -04:00
Patrick Hilhorst
1f7fc09176
make-derivation: use a more descriptive assert message
As suggested by @Profpatsch
2018-10-30 14:33:14 +01:00
Léo Gaspard
02e1f00ffd
dovecot, opensmtpd: add link to test in meta.tests
Rationale
---------

Currently, tests are hard to discover. For instance, someone updating
`dovecot` might not notice that the interaction of `dovecot` with
`opensmtpd` is handled in the `opensmtpd.nix` test.

And even for someone updating `opensmtpd`, it requires manual work to go
check in `nixos/tests` whether there is actually a test, especially
given not so many packages in `nixpkgs` have tests and this is thus most
of the time useless.

Finally, for the reviewer, it is much easier to check that the “Tested
via one or more NixOS test(s)” has been checked if the file modified
already includes the list of relevant tests.

Implementation
--------------

Currently, this commit only adds the metadata in the package. Each
element of the `meta.tests` attribute is a derivation that, when it
builds successfully, means the test has passed (ie. following the same
convention as NixOS tests).

Future Work
-----------

In the future, the tools could be made aware of this `meta.tests`
attribute, and for instance a `--with-tests` could be added to
`nix-build` so that it also builds all the tests. Or a `--without-tests`
to build without all the tests. @Profpatsch described in his NixCon talk
such systems.

Another thing that would help in the future would be the possibility to
reasonably easily have cross-derivation nix tests without the whole
NixOS VM stack. @7c6f434c already proposed such a system.

This RFC currently handles none of these concerns. Only the addition of
`meta.tests` as metadata to be used by maintainers to remember to run
relevant tests.
2018-10-30 21:31:39 +09:00
Patrick Hilhorst
c7e026bec4
make-derivation: use lib.assertMsg
As suggested by @Profpatsch
2018-10-29 18:25:59 +01:00
Patrick Hilhorst
5be927db14
make-derivation: use ? instead of builtins.hasAttr
As suggested by @edolstra
2018-10-29 15:17:13 +01:00
Patrick Hilhorst
2962f94fec
make-derivation: add check that the name is consistent with pname and version 2018-10-29 14:58:12 +01:00
Patrick Hilhorst
149a55eca7
make-derivation: get position info from version 2018-10-29 14:51:22 +01:00
Patrick Hilhorst
efca8b4b97
make-derivation: use pname-version as default name if both are present 2018-10-29 14:51:08 +01:00
Matthew Bauer
b3041b4455 make-derivation: set CMAKE_SYSTEM_* when cross compiling
Uses uname data to find what to set these variables:

- CMAKE_SYSTEM_NAME
- CMAKE_SYSTEM_PROCESSOR
- CMAKE_SYSTEM_VERSION
- CMAKE_HOST_SYSTEM_NAME
- CMAKE_HOST_SYSTEM_PROCESSOR
- CMAKE_HOST_SYSTEM_VERSION
2018-10-16 21:50:37 -05:00
Yegor Timoshenko
cd0c8739d7
Merge pull request #37600 from abbradar/impureusenative
impureUseNativeOptimizations: add stdenv adapter
2018-10-13 14:09:55 +00:00
Matthew Bauer
93834fe194
Merge pull request #47230 from bhipple/fix/licenses
Remove dead code from stdenv check-meta license logic
2018-10-05 22:58:27 -05:00
John Ericson
f49ca01c50 Revert "stdenv: partial revert of f2bb59e"
This reverts commit 607063f61b.
2018-09-26 14:47:16 -04:00
Matthew Bauer
607063f61b stdenv: partial revert of f2bb59e
/cc @Ericson2314

PR was https://github.com/NixOS/nixpkgs/pull/46857

This line broke MacOS cross compilation. paxctl cannot be built on
macOS. Maybe it can be fixed, but no reason to break things
unnecessarily.

Regardless, you definitely need to be more careful about backporting.
I think it’s fine to move fast and break things on master but
with release-18.09 we should be more careful. Something like more
automated testing for cross compilation would also be
helpful (hopefully even making it block).

(cherry picked from commit f9c4075873)
2018-09-26 11:13:22 -04:00